The aim of this proposal is the rational design of protease inhibitors based on knowledge of the catalytic mechanisms of the target enzymes. These inhibitors will then be used to elucidate the physiological roles of specific proteases in a limited number of in vivo model systems. The zinc metalloproteases angiotensin converting enzyme and collagenase are the target enzymes. Phosphorus-containing transition state inhibitors of converting enzyme which have already been developed will be modified in order to improve their stability in vivo. Phenethylphosphonylalanylproline, the most active compound prepared so far, will be alkylated on one phosphorus oxygen to give inactive phenethyl-(Ophenyl)-phosphonylalanylproline which is stable to low pH but will spontaneously hydrolyze at pH 7.4 to the active parent compound. It will be tested for oral activity in hypertensive rats. Other transition state analogs based on arsenic and silicon instead of phosphorus will be investigated in vitro. The investigation of another class of converting enzyme inhibitors, suicide substrates, will be continued. 3,4-Epoxy-2-methyl-butanoylproline was found not to be a suicide inhibitor of converting enzyme. This compound will be further investigated to determine whether the enzyme catalyzes the opening of the expoxide by the addition of water instead of an enzyme-bound nucleophile, which would explain why it is not a suicide inhibitor. Phosphorus containing inhibitors of both bacterial and vertebrate collagenase will be prepared until inhibitors with Ki's in the 10-100 nM range are found. Both converting enzyme and vertebrate collegenase inhibitors will be tested for activity against tumor induced angiogenesis in two models, the rabbit cornea and the chorioallantoic membrane of the fertilized chicken egg. If these inhibitors are antiangiogenic, it would be the first indication that converting enzyme and vertebrate collagenase are required in the process of angiogenesis. Inhibitors active in these models will be tested against tumor growth in vivo. If the sequences around the susceptible bonds in the different types of collagens become available, then inhibitors against the vertebrate collagenases specific for each collagen could be prepared. Thus the metastasis of tumors through basement membrane collagen could be inhibited without affecting bone remodelling, for example.